debug=True in production gives anyone a Python console on your server. A hard-coded SECRET_KEY lets anyone forge sessions. This skill finds those and ten other Flask mistakes before you deploy, and tells you how to fix each one.

Your agent runs a scanner on your Python files and Jinja templates. It reads Python with Python's own parser, follows values from request.args, request.form, request.json and URL variables through your code, and knows the safe patterns: environment variables, secure_filename, safe_join, url_for, parameterised SQL and number URL converters. It reads .gitignore so secrets in ignored config files are not reported. It never runs your app. It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach. Secrets are never printed.

What it checks

BLOCK (fix before shipping)

• F01: debug mode on (app.run(debug=True), DEBUG = True)

• F02: a hard-coded SECRET_KEY, JWT key, password, token or API key

• F03: render_template_string with a template that is not fixed (template injection)

• F04: SQL built with f-strings, %, .format() or + and run with execute or text()

• F05: send_file or open with a path from the request

REVIEW (check before release)

• F11: CORS with credentials and no origin list

• F12: Markup() on request data, |safe or autoescape off in templates

• F13: session or remember cookies without Secure or HttpOnly

• F14: CSRF turned off or views exempted

• F15: redirect to a URL from the request

• F16: uploads saved with the user's file name

• F17: app.run(host='0.0.0.0')

What's included

• The scanner (Python 3.9+, standard library only)

• A rules guide with the reason and safer approach for each rule

• A risky and a safe example Flask app, and the real report the scanner produces on the risky one

Limits

It follows values within one file. Extensions' own settings, your WSGI server, reverse proxy and dependencies are not checked. NOT FLAGGED is not a guarantee.

Contents

One ZIP file with 13 files: SKILL.md, the scanner in scripts/ (3 Python files), references/rules.md (every rule with its reason and safer approach), and examples/ (7 example files plus report.md, the real scanner output on the risky example).

How to use: This is an agent skill: a folder with a SKILL.md file and Python scripts that an AI coding agent such as Claude Code, Cursor, or Codex CLI can run. Unzip it into your agent's skills folder (for Claude Code: ~/.claude/skills/), then ask your agent to check your project. The scripts also run on their own with Python 3.9 or newer, with nothing to install. They read your files only: no network access, and nothing is sent anywhere.